Sign in with
Sign up | Sign in
Your question

Don Maddox interviews the "Stars"

Last response: in Video Games
Share
Anonymous
August 8, 2005 4:18:36 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Hi,

The StrategyZone's Don Maddox interviews Dave O'Connor, John Tiller,
Gregor Whiley, Mark H. Walker and Patrick Proctor on the current state
of wargaming and what they think the future will bring.

http://www.strategyzoneonline.com/articles.php?p=507&pa...

No surprises if you've been following "the scene" for some time, but
when you read the contrasting views in one article you sometimes must
start to doubt they're all in the same business.

Greetz,

Eddy Sterckx
Anonymous
August 8, 2005 6:46:57 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

JP schreef:

> Where's the Major ?

Well, I for one would like to get his view on the issue of the
probability/outlook wrt future funding of military simulations by the
various Dod's of this planet.

The US military have certainly embraced games as a pr and recruiting
tool with the latest offering from the Navy even being educational on
marine wildlife but is there any money left for training & simulation
programs ?

And what about the non-US countries ? Nato certainly is a joke wrt this
as every country has developed their own in-house training tools (or
bought the Israeli stuff) instead of pooling their efforts or adapting
commercially available simulations.

Greetz,

Eddy Sterckx
August 8, 2005 8:36:09 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

<eddysterckx@hotmail.com> wrote in message
news:1123485516.856462.219240@g14g2000cwa.googlegroups.com...
> Hi,
>
> The StrategyZone's Don Maddox interviews Dave O'Connor, John Tiller,
> Gregor Whiley, Mark H. Walker and Patrick Proctor on the current state
> of wargaming and what they think the future will bring.
>
> http://www.strategyzoneonline.com/articles.php?p=507&pa...
>
> No surprises if you've been following "the scene" for some time, but
> when you read the contrasting views in one article you sometimes must
> start to doubt they're all in the same business.
>
> Greetz,
>
> Eddy Sterckx



Where's the Major ?
Related resources
Anonymous
August 8, 2005 10:59:37 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Giftzwerg schreef:

> All of the "advances" in games like SIM CITY or DOOM are in terms of
> *graphics*. If you look at actual *gameplay*, DOOM III is nearly
> indistinguishable from the original DOOM; you run around, shooting
> critters that pop out at you, trying to find The McGuffin. In that
> sense, there have been *no advances at all*!

<scratching my head> don't know how I missed that one - razor sharp
observation.

> And that brings us directly to the obvious answer. It isn't that
> "existing [wargame] designs are so good that there is no need to
> evolve," but that the only meaningful advance we can bring to wargames
> is in terms of AI. Even some modern features of wargames like
> simulation of command levels by removing the ability to micromanage
> subordinate units, or WEGO or PCT game structures, are entirely
> predicated on advancing the state of AI development.

You drank some powerfull stuff this morning :)  - I have to agree, even
*really* innovative stuff like team multi-player or convincing
diplomacy are inherently AI advancements.

Greetz,

Eddy Sterckx
Anonymous
August 8, 2005 1:16:24 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

In article <1123485516.856462.219240@g14g2000cwa.googlegroups.com>,
eddysterckx@hotmail.com says...

> No surprises if you've been following "the scene" for some time, but
> when you read the contrasting views in one article you sometimes must
> start to doubt they're all in the same business.

"So what is the current state of computer wargaming? It is abundantly
clear that wargames have not kept pace with the impressive improvements
seen over the last decade in other genres. One only has to glance at the
difference between the original Sim City and Sim City 4, or the original
Doom and Doom III, to see just how far PC gaming has come from its early
days. But during that same period wargames have generally remained the
same. Are we to believe that this is because existing designs are so
good that there is no need to evolve?"
- Don Maddox

This is Maddox's main question, and - except for O'Connor, who touched
on it in passing - I'm appalled that the obvious answer doesn't pop
right out of all of these designer's mouths.

All of the "advances" in games like SIM CITY or DOOM are in terms of
*graphics*. If you look at actual *gameplay*, DOOM III is nearly
indistinguishable from the original DOOM; you run around, shooting
critters that pop out at you, trying to find The McGuffin. In that
sense, there have been *no advances at all*!

And that brings us directly to the obvious answer. It isn't that
"existing [wargame] designs are so good that there is no need to
evolve," but that the only meaningful advance we can bring to wargames
is in terms of AI. Even some modern features of wargames like
simulation of command levels by removing the ability to micromanage
subordinate units, or WEGO or PCT game structures, are entirely
predicated on advancing the state of AI development.

What I mean here is that advancing the state of wargaming depends on the
most difficult to achieve advances in computing technology. Compared to
that, smaller pixels in DOOM III were almost trivial. Thus it's not
"abundantly clear" to me that wargame development is lagging - only that
the other genres had such easy advances over the last 20 years that a
superficial reading of history leads us to a terrible error in
judgement.

--
Giftzwerg
***
"The most despicable misuse of terminology, however, occurs when
Leftists call the Saddamites and foreign jihadists 'the resistance.'
What an example of moral inversion! For the fact is, paramilitary death
squads are attacking the Iraqi people. And those who oppose the killers
- the Iraqi police and National Guardsmen, members of the Allawi
government, people like Nour - they are the "resistance."
- Stephen Vincent
(Journalist murdered by terrorists, 8/2/2005)
Anonymous
August 8, 2005 2:19:30 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Giftzwerg wrote:
> In article <1123485516.856462.219240@g14g2000cwa.googlegroups.com>,
> eddysterckx@hotmail.com says...
>
>
>>No surprises if you've been following "the scene" for some time, but
>>when you read the contrasting views in one article you sometimes must
>>start to doubt they're all in the same business.
>
>
> "So what is the current state of computer wargaming? It is abundantly
> clear that wargames have not kept pace with the impressive improvements
> seen over the last decade in other genres. One only has to glance at the
> difference between the original Sim City and Sim City 4, or the original
> Doom and Doom III, to see just how far PC gaming has come from its early
> days. But during that same period wargames have generally remained the
> same. Are we to believe that this is because existing designs are so
> good that there is no need to evolve?"
> - Don Maddox
>
> This is Maddox's main question, and - except for O'Connor, who touched
> on it in passing - I'm appalled that the obvious answer doesn't pop
> right out of all of these designer's mouths.
>
> All of the "advances" in games like SIM CITY or DOOM are in terms of
> *graphics*. If you look at actual *gameplay*, DOOM III is nearly
> indistinguishable from the original DOOM; you run around, shooting
> critters that pop out at you, trying to find The McGuffin. In that
> sense, there have been *no advances at all*!
>
> And that brings us directly to the obvious answer. It isn't that
> "existing [wargame] designs are so good that there is no need to
> evolve," but that the only meaningful advance we can bring to wargames
> is in terms of AI. Even some modern features of wargames like
> simulation of command levels by removing the ability to micromanage
> subordinate units, or WEGO or PCT game structures, are entirely
> predicated on advancing the state of AI development.
>
> What I mean here is that advancing the state of wargaming depends on the
> most difficult to achieve advances in computing technology. Compared to
> that, smaller pixels in DOOM III were almost trivial. Thus it's not
> "abundantly clear" to me that wargame development is lagging - only that
> the other genres had such easy advances over the last 20 years that a
> superficial reading of history leads us to a terrible error in
> judgement.
>

It's true that other genres have advanced in terms of graphics but also
in many areas as well. The RPG now is light years more advanced than it
was ten years ago. Why, is the question facing wargamers? The answer is
in my opinion, a commitment to AI, an area that the wargame developers
have shown only a passing interest. Where is this supposed game play
developments in wargames when the AI remains as crippled as ever. Also,
the wargame developers have shown little interest in embracing new
technology as well. Sure we get some nice graphics like in TOAW but
that's about it. Fundamentally, how is my copy of new Kursk '43 a huge
leap from the 80's kampfgruppen?

Alan
Anonymous
August 8, 2005 3:06:43 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

In article <WRJJe.853$ah4.679@lakeread05>, Alahorn@cox.net says...

> It's true that other genres have advanced in terms of graphics but also
> in many areas as well.

Which ones? Exactly.

> The RPG now is light years more advanced than it
> was ten years ago.

How so? Except for better graphics and increased support for complex
script environments, I'm not sure there have been *any* advances in the
RPG genre since I bought my copy of FALLOUT way-back-when.

> Why, is the question facing wargamers? The answer is
> in my opinion, a commitment to AI, an area that the wargame developers
> have shown only a passing interest.

I'm not sure what you're saying here. My point is that wargames - like
no other genre, really - are fundamentally dependant upon AI advances to
be thought of as "evolutionary." And AI advances aren't easy ones, like
graphics, where a new generation of "AI cards" appears and allows you to
double the "smartness" of your PO programming.

So I'm not sure it's a question of "wargame programmers aren't
interested," but "wargame programmers have to do all the AI work
themselves."

> Where is this supposed game play
> developments in wargames when the AI remains as crippled as ever.

This is one of those points I wish I'd made.

....wait...

> Also,
> the wargame developers have shown little interest in embracing new
> technology as well.

Like what? DVD burners? SAN? Fingerprint scans?



--
Giftzwerg
***
"The most despicable misuse of terminology, however, occurs when
Leftists call the Saddamites and foreign jihadists 'the resistance.'
What an example of moral inversion! For the fact is, paramilitary death
squads are attacking the Iraqi people. And those who oppose the killers
- the Iraqi police and National Guardsmen, members of the Allawi
government, people like Nour - they are the "resistance."
- Stephen Vincent
(Journalist murdered by terrorists, 8/2/2005)
Anonymous
August 8, 2005 6:09:49 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Alan H schreef:

> I have to question how AI advances can't anything but easy. We have a
> computer that can process information 1000s of times faster than before,
> the AI is just a set of actions based on conditions. In terms of raw
> processing power alone the technology exists to process a massive amount
> of these conditions.

AI development was never a problem of processing known conditions -
that's scripting - it's about how to cater for the unexpected, for the
strategic surprise.

Take such a simple thing as a frontline - it takes a grognard eye about
a second to spot a hole in it, to see a unit that is over-extended, to
see the schwerpunkt of the attack. Now try to come up with algorythms
to determine these concepts. Hard,but do-able. Now try to cater for
special cases : surrounded units, enemy units appearing in your flank,
even determining what your flank actually is. Determining when it's
time to pull back etc.

> A RPG game can be created with multitudes of
> "complex script environments" that shows that this can be done. Yet, the
> wargame AI still sits there in a thorizne induced coma or attacks
> piecemeal like an impatient two year old. How can we have in BG2 for
> example plots upon subplots all script drive and yet my Tiller AI is
> just as retarded now in France '40 as when I bought Normandy '44.

Simply because it's the same retarted AI as featured in the first
BattleGround game 10 years ago ?

Without even mentioning very good AI's as those made by SSG (Battles in
Normandy/Italy and Panther Games (HTTR) you'll be hard-pressed to even
find a game with a *worse* AI then those you mentioned above.

> I'm thinking of Big Blue and how it took on Kasparov. All that computer
> did was examine the possible moves with brute force crunch the numbers,
> nothing fancy about that AI. Today, I have a P4 3ghz. Yesterday I had a
> 386 16mhz. The ability to double the "smartness" of the AI exists. In
> general my point is that the wargame developers/programmers have yet to
> take advantage of the current state of technology either processing
> power or graphically.

The real problem is that you've been playing the wrong games - sorry,
but that's the way it is. You just can't play Tiller games and then
complain about graphics or AI - these games are tailor-made for PBEM
play and for people who don't have the time/patience to learn a new
game every couple of months.

If you want good AI opponents and/or good graphics check out what's
beyond the HPS border - from the games you mentioned I gather you like
operational level games so the "Battles in ..." and HTTR games are
maybe what you should check out. And their graphics are better too.

Greetz,

Eddy Sterckx
Anonymous
August 8, 2005 7:54:58 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Giftzwerg wrote:
> In article <WRJJe.853$ah4.679@lakeread05>, Alahorn@cox.net says...
>
>
>>It's true that other genres have advanced in terms of graphics but also
>>in many areas as well.
>
>
> Which ones? Exactly.
See Below

>
>
>>The RPG now is light years more advanced than it
>>was ten years ago.
>
>
> How so? Except for better graphics and increased support for complex
> script environments, I'm not sure there have been *any* advances in the
> RPG genre since I bought my copy of FALLOUT way-back-when.

Just compare fallout to wasteland or compare Morrowind to Ultima. The
progress is beyond just graphics they have used the new horsepower of
the Pentium to create a whole interactive world. Sure it might be the
same 1's and 0's but they have changed with the progress of technology
to adpapt and offer something new every time.

>
>
>>Why, is the question facing wargamers? The answer is
>>in my opinion, a commitment to AI, an area that the wargame developers
>>have shown only a passing interest.
>
>
> I'm not sure what you're saying here. My point is that wargames - like
> no other genre, really - are fundamentally dependant upon AI advances to
> be thought of as "evolutionary." And AI advances aren't easy ones, like
> graphics, where a new generation of "AI cards" appears and allows you to
> double the "smartness" of your PO programming.
>
> So I'm not sure it's a question of "wargame programmers aren't
> interested," but "wargame programmers have to do all the AI work

I have to question how AI advances can't anything but easy. We have a
computer that can process information 1000s of times faster than before,
the AI is just a set of actions based on conditions. In terms of raw
processing power alone the technology exists to process a massive amount
of these conditions. A RPG game can be created with multitudes of
"complex script environments" that shows that this can be done. Yet, the
wargame AI still sits there in a thorizne induced coma or attacks
piecemeal like an impatient two year old. How can we have in BG2 for
example plots upon subplots all script drive and yet my Tiller AI is
just as retarded now in France '40 as when I bought Normandy '44.

I'm thinking of Big Blue and how it took on Kasparov. All that computer
did was examine the possible moves with brute force crunch the numbers,
nothing fancy about that AI. Today, I have a P4 3ghz. Yesterday I had a
386 16mhz. The ability to double the "smartness" of the AI exists. In
general my point is that the wargame developers/programmers have yet to
take advantage of the current state of technology either processing
power or graphically.


> themselves."
>
>
>>Where is this supposed game play
>>developments in wargames when the AI remains as crippled as ever.
>
>
> This is one of those points I wish I'd made.
>
> ...wait...
>
>
>>Also,
>>the wargame developers have shown little interest in embracing new
>>technology as well.
>
>
> Like what? DVD burners? SAN? Fingerprint scans?
>
Ignore that last comment.

>

Alan
Anonymous
August 8, 2005 8:14:58 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> Where's the Major ?

I respectfully declined to participate. I thought that the proposed
interview questions should be put to a group of knowledgeable wargamers
(paying customers) rather than to a group of developers. Perhaps then
followed by responses by developers.

That being said, and after a quick read, I didn't see anything by Dave
O¹Connor, John Tiller, Gregor Whiley, Mark H. Walker, or Patrick Proctor
that I disagreed with.

Best regards, Major H.
tacops@mac.com
http://www.battlefront.com/
Anonymous
August 9, 2005 12:06:21 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

In article <pMOJe.861$ah4.833@lakeread05>, Alahorn@cox.net says...

> > How so? Except for better graphics and increased support for complex
> > script environments, I'm not sure there have been *any* advances in the
> > RPG genre since I bought my copy of FALLOUT way-back-when.
>
> Just compare fallout to wasteland or compare Morrowind to Ultima. The
> progress is beyond just graphics they have used the new horsepower of
> the Pentium to create a whole interactive world. Sure it might be the
> same 1's and 0's but they have changed with the progress of technology
> to adpapt and offer something new every time.

Sure, but the "whole interactive world" is really just the same old
interactive world dressed up with more of everything; scripts,
characters, settings, etc. FALLOUT *is* WASTELAND, except bigger,
better, and prettier; more weapons, more levels, more story, more depth,
more quests ... more of everything. Gameplay-wise, though, I'm not sure
it's possible to tell the difference.

> > So I'm not sure it's a question of "wargame programmers aren't
> > interested," but "wargame programmers have to do all the AI work
>
> I have to question how AI advances can't anything but easy. We have a
> computer that can process information 1000s of times faster than before,
> the AI is just a set of actions based on conditions. In terms of raw
> processing power alone the technology exists to process a massive amount
> of these conditions. A RPG game can be created with multitudes of
> "complex script environments" that shows that this can be done. Yet, the
> wargame AI still sits there in a thorizne induced coma or attacks
> piecemeal like an impatient two year old.

Well ... let me back up a moment.

Clearly, there are *some* designers out there who just aren't interested
- exactly as you assert; you've got me cold with the "Tiller" reference.

But there are some fairly smart AIs out there, at least of late, so I'm
not sure it's fair to tar the entire industry with Tiller and the other
scenario-pack-wranglers.

> How can we have in BG2 for
> example plots upon subplots all script drive and yet my Tiller AI is
> just as retarded now in France '40 as when I bought Normandy '44.

In a word, *replayability*. People play wargame scenarios over and over
- while they tend to play RPG's *once*. That means you can get away
with depending almost totally on scripts for RPG coding, in a way that
you simply cannot with a wargame. I mean, I could create a *completely*
scripted wargame scenario that would knock your socks off, smarts-wise
.... once. And the second time you played it, you'd be bored to tears.

In fact, I just replayed FALLOUT2, and it was fairly enjoyable; the
intervening six or seven years served to pretty much obliterate anything
but a general idea of what was going on. But I've played HTTR
*continuously for almost two years*! If Panther had relied on
scripting, it would have interested me for about two months.

[And there are some hybrid approaches, IE SSG's BATTLES IN ... series,
where multiple scripts can serve as a jumping off point for AI
processing.]

> I'm thinking of Big Blue and how it took on Kasparov. All that computer
> did was examine the possible moves with brute force crunch the numbers,
> nothing fancy about that AI. Today, I have a P4 3ghz. Yesterday I had a
> 386 16mhz.

Yeah, but chess is *trivial* compared with a wargame if you're
experimenting with brute force algorithms. Deep Blue was working with
exactly 16 maneuver units on each side (of exactly six types), on a
board with exactly 64 areas - and they needed a phenomenally powerful
supercomputer with specially designed "chess chips" that were hard-coded
for performance in this one selected area.

How many units in a game of, say, RUSSO-GERMAN WAR? How many different
types of units? How many hexes? I'd be willing to bet (based on a
vague suspicion only...) that the raw number-crunching for a middlin'
wargame is five orders of magnitude more difficult that a simple little
game of chess.

[On the other hand, they had to beat Kasparov - and CHESSMASTER 9000 has
been kicking my ass quite handily since it was CHESSMASTER 2000...]

> The ability to double the "smartness" of the AI exists. In
> general my point is that the wargame developers/programmers have yet to
> take advantage of the current state of technology either processing
> power or graphically.

<shrug>

Not being an AI programmer, I really don't have a feel for this, either
way. But I'm pretty sure that programming a digital general to be as
smart as a human being is *substantially* more difficult than, say,
programming a digital airplane to look and fly like a real one. I mean,
we've had flight simulators that work *exactly* like a real airplane for
more than a decade now - but human pilots are still sitting firmly in
airline cockpits around the world.

--
Giftzwerg
***
"The most despicable misuse of terminology, however, occurs when
Leftists call the Saddamites and foreign jihadists 'the resistance.'
What an example of moral inversion! For the fact is, paramilitary death
squads are attacking the Iraqi people. And those who oppose the killers
- the Iraqi police and National Guardsmen, members of the Allawi
government, people like Nour - they are the "resistance."
- Stephen Vincent
(Journalist murdered by terrorists, 8/2/2005)
Anonymous
August 9, 2005 2:05:14 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> I respectfully declined to participate. I thought that the proposed
> interview questions should be put to a group of knowledgeable wargamers
> (paying customers) rather than to a group of developers. Perhaps then
> followed by responses by developers.

I fully agree. Only a couple of weeks ago, discussing the comic-book market
situation, I proposed to my fellow artists that the planning for the next
year or two should include some very honest feedback from the paying
public - since it is really, *really* too easy to build around yourself the
"perfect vision" of the market only to complain that the P.P. (paying
public) wasn't ready for the "vision" when something dire happens. For some
reason my proposal was quickly dropped (well, it included showing such
feedback to the Editor, khof, ahem...) but it would be interesting to
"interview" some P.P. (player personalities) like Giftzy [1] or Eddy about
the current state of wargaming the same way developers are interviewed.

[1] Put as a condition that the word "cookiecutter" cannot be used in the
"pars destruens" of your answers, and I believe that, for once, we will see
Gifty gasping for air ^____^
Anonymous
August 9, 2005 9:43:52 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Arjuna schreef:

> So what it boils down to is a question of priorities. That's why we at
> Panther Games are going to be focusing on just one or two new feature
> per title. ( I know COTA will have a swag of new features. You just
> can't hold some people back ;)  )

<innocently> Really ? </innocently>

> So what would you see as the top ten new features you would like to see
> in future releases?

ok, for the benefit of the discussion I'll put my top 5 of ideas in
here - Arjuna allready knows - and ignores :)  them - but maybe they'll
trigger a reaction or two.

1) PBEM - In the current engine when you load a previously saved game
the side you play is fixed as the one you've originally chosen. But
what if the engine could switch sides during game-load ?

The pbem setup might look something like this :

- game is played without orders delay
- player A gives his orders and lets the game run for half an hour,
the player B forces are handled by the AI at this point. After that
half hour the game is saved and mailed to the opponent.
- player B loads the game - maybe auto-swich sides here - reviews the
situation, and can give orders for the next 30 minutes while the AI
controls the other side.

This way there's an automatic order delay - 30 minutes in which you
can't issue orders yourself but must rely on your subordinates i.e. the
AI

I've used 30 minutes as an example, but if you really want to go wild
you can let the players determine it amongst themselves how long a
"turn" lasts.

2) a Bid System - Strategic Command uses a bid system in it's
scenario's to finely balance them for multi-player. If you happen to
think a certain scenario is unbalanced you are kindly requested to put
your money where your mouth is - or more appropriatly : put some VP's
on the line.

Here is how it could work : In the eMail exchange before an online or
pbem game both players make a bid in VP's to play the Germans, the
highest bid gets to play them. If you think the Germans don't stand a
chance in this scenario enter a negative bid. Works the same.

In the scenario-setup screen : put in an editable field where players
can put in the bid number - this is the amount of VP's that will get
deducted from the German total at game's end. A negative highest bid
gets added to the total. Normal victory determination from there on.

3) player controllable dummy units that are seen by the enemy player/ai
as real units. Use them to make a feint at one objective, but go for
another. Would add a lot of spice to some battles where the current FoW
just isn't good enough. And the AI would get to do that as well of
course.

4) Air recce - I even had to *cite* books before they would believe
corps level commanders had access to air recce resources they could
controll themselves. It's a tough crowd over there in the beta-forum
.... :) 

5) letting the gamer place (virtual) markings on the map - right double
click or such brings up a menu "set mark" and "delete mark" or simply
setting and deleting without the menu.

A hotkey instantly centers the map at the current zoomlevel to the next
marking in the runtime list.

This would allow quick switching between various hotspots on the map as
the current combo of using the minimap and scrolling doesn't really
work for me at the more detailed zoom-levels

Greetz,

Eddy Sterckx
Anonymous
August 9, 2005 12:07:56 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Some disconnected thoughts on an interesting article:

SZO: "Do you think computer wargaming has fallen behind other gaming
genres?"

I'm not sure how useful it is to view the wargames industry as a single
entity anymore. There are companies like Battlefront, Panther and
MadMinute
implimenting exciting, progressive ideas that move the hobby forward and
attract new blood. There are designers and designs so fossilized they
wouldn't know innovation or elegance if they were bitten on the bum by
them. Chalk and cheese. Two factions with little in common besides an
appetite for realism.

Patrick Proctor, Prosim: "For the industry to innovate, there has to be
incentive to innovate."

Could there really be a bigger, more obvious incentive to innovate than
Charles Moylan's gold-plated Kubelwagen or Steve Grammont's
diamond-encrusted Kettenkrad? The most commercially successful wargame of
the past five years is also the most innovative. There has to be a lesson
there somewhere.

Mark H. Walker: "I think that we would all kill for games that look like
CDV’s stuff, yet play like Steel Panthers or Lock ‘n Load, but those type
of graphics cost money, big money."

A weak excuse? Get your visuals made somewhere relatively cheap like the
former SU, employ a few talented amateurs on a profit-share basis, create
a game with simple yet stylish models and textures that can be modded
easily... there are ways of producing a great-looking game without
massive investment.

Gregor Whiley, SSG
"Computer wargames have never been a mass market genre"

Tell that to Creative Assembly. What are the Total War series if they're
not user-friendly lite wargames with great visuals and highly developed
campaigns. Theres nothing integral to wargaming that means it has to be a
niche sector.
Anonymous
August 9, 2005 12:09:41 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Giftzwerg schreef:

> I dunno about this one. My "Potential For Gamey Abuses" meter is pegged
> hard at the "red" end of the scale, and a dozen alarms are blaring.
>
> *User-exploitable* Fog-Of-War, that can be used in a systematic way?

Yup, I don't need to tell you that I'm alone on this in the beta-forum
too - too many sensible heads over there, while I think it was an
integral part of some campaigns, most notably the Bulge (Tigers !
Tigers ! everywhere !) and the North African campaign. Not to mention
the Ghost Army assembled in Kent ready to invade the Pas de Calais area
in 1944 ...

Gamey : not if done correctly - time-limited to the starting hours of a
scenario or limited to a few units. Not a major game decider, but ...
Never mind, I'm probably the only one in the universe who thinks it's a
good idea, but that won't stop me promoting it :) 

Greetz,

Eddy Sterckx
Anonymous
August 9, 2005 1:07:45 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

In article <1123591432.857778.294640@g49g2000cwa.googlegroups.com>,
eddysterckx@hotmail.com says...

> > So what would you see as the top ten new features you would like to see
> > in future releases?

> 3) player controllable dummy units that are seen by the enemy player/ai
> as real units. Use them to make a feint at one objective, but go for
> another. Would add a lot of spice to some battles where the current FoW
> just isn't good enough. And the AI would get to do that as well of
> course.

I dunno about this one. My "Potential For Gamey Abuses" meter is pegged
hard at the "red" end of the scale, and a dozen alarms are blaring.

*User-exploitable* Fog-Of-War, that can be used in a systematic way?

--
Giftzwerg
***
"The most despicable misuse of terminology, however, occurs when
Leftists call the Saddamites and foreign jihadists 'the resistance.'
What an example of moral inversion! For the fact is, paramilitary death
squads are attacking the Iraqi people. And those who oppose the killers
- the Iraqi police and National Guardsmen, members of the Allawi
government, people like Nour - they are the "resistance."
- Stephen Vincent
(Journalist murdered by terrorists, 8/2/2005)
Anonymous
August 9, 2005 3:07:39 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> The real problem is that you've been playing the wrong games - sorry,
> but that's the way it is. You just can't play Tiller games and then
> complain about graphics or AI - these games are tailor-made for PBEM
> play and for people who don't have the time/patience to learn a new
> game every couple of months.

Which suggests to me that it is also necessary to define exactly what is
meant by wargame "evolution" or "progress".

Tiller fans are obviously quite happy with the system they've got and simply
want that system to present them with more and different military-historical
challenges.

Others insist that progress is about developing better systems to re-examine
tried-and-true situations along the lines of "it may be 'just another Bulge
game' but it uses better algorithms that generate a whole new gaming
experience" (or whatever :-) ).

It seems to me that one version of "progress" does involve increasing
players' sense that they are "commanding forces" rather than "playing a
game" and that this does involves more sophisticated AI routines (for the
player's side, not the enemy's) if it's to be done properly. OTOH, user
interface improvements (the information a player is given and how it is
presented) can also make a difference. A wargame's popularity may be less
graphics-dependent (among its target market) than FPS/RTS/RPG but a good
user interface can enhance a wargame's atmosphere while a bad one can spoil
the effect of otherwise brilliant modelling and coding (TOPS/PITS?).

Cheers,

Andy
Anonymous
August 9, 2005 3:07:40 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Arjuna wrote:
>
> Would you the paying customers be prepared to pay $100 per game to pay
> for that? I don't think so. But hey tell me you will and we'll jump for
> joy :) )

Well, "War in the Pacific" cost me $80, so we're getting close! :-(

>
> So what would you see as the top ten new features you would like to see
> in future releases?
>

1-9: Better AI
10: Better UI


;-)
Anonymous
August 9, 2005 9:29:06 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

>3) player controllable dummy units that are seen by the enemy player/ai
>as real units. Use them to make a feint at one objective, but go for
>another. Would add a lot of spice to some battles where the current FoW
>just isn't good enough. And the AI would get to do that as well of
>course.

LOL

Oh don't you just love those throw away lines like "And the AI would
get to do that as well of course." ;) 

Sure thing Eddy...ready by dinner mate. ;) 

four hours later...Eddy got any ideas how, when and why the AI should
do this and when and why it should not do it? If not, don't worry, I'll
just use a random die roll. That should add to the realism, hey? ;) 

And don't worry about responding Eddy that's retorical. ;)  :) )
Anonymous
August 10, 2005 4:05:03 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> we've had flight simulators that work *exactly* like a real airplane for
> more than a decade now - but human pilots are still sitting firmly in
> airline cockpits around the world.

This is really only because it's felt the paying, air-travelling public
would balk at flying in a plane with no pilot.

I believe that automating a flight from before takeoff 'til after landing is
now a relatively trivial thing to do but no one's yet prepared to do away
with the pilot just in case "something" does go wrong.

I understand from mates in the civil aviation business that, frequently
these days, the modern pilot is actually the least reliable "system" on
board!

Andy
Anonymous
August 10, 2005 4:05:04 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

In article <dda65k$so1$1@lust.ihug.co.nz>, andybrn@somewhere.in.nz
says...

> > we've had flight simulators that work *exactly* like a real airplane for
> > more than a decade now - but human pilots are still sitting firmly in
> > airline cockpits around the world.
>
> This is really only because it's felt the paying, air-travelling public
> would balk at flying in a plane with no pilot.

My point, though, is that flight simulators are pretty much a mature
technology; until somebody invents holographic direct-inputs to the
brain, they're really as good as they're ever going to get *or need to
be*. But AI pilots have a really, really, really long way to go before
anyone - the public *or the guys who own that $300,000,000 airplane* -
is going to entrust them with sole responsibility for the flight.

That's the issue here. All a flight-sim designer has to do is build
technology that's been around for 20 years into a nice looking package
for the PC. A wargame designer has to build (or fake) something that's
eluded pretty much everyone, including the MIT AI lab. Arthur C. Clarke
(no dummy, him...) thought we'd have human-level AI by 2001[1]; I'm not
sure we'll have it in 2101.

> I believe that automating a flight from before takeoff 'til after landing is
> now a relatively trivial thing to do but no one's yet prepared to do away
> with the pilot just in case "something" does go wrong.

They haven't eliminated railroad engineers, either, although I can't
imagine anything more eye-glazingly tedious than MICROSOFT TRAIN
SIMULATOR.

> I understand from mates in the civil aviation business that, frequently
> these days, the modern pilot is actually the least reliable "system" on
> board!

....until an engine comes apart, and takes out all the hydraulics, and
the autopilot's only response is to helplessly go, "Beeeep, Beeeep,
Beeeep."

Then that bleary-eyed jamoche is worth every nickel of the pittance they
pay him.


[1] "You'll find that rather difficult without your space helmet, Dave."

--
Giftzwerg
***
"The most despicable misuse of terminology, however, occurs when
Leftists call the Saddamites and foreign jihadists 'the resistance.'
What an example of moral inversion! For the fact is, paramilitary death
squads are attacking the Iraqi people. And those who oppose the killers
- the Iraqi police and National Guardsmen, members of the Allawi
government, people like Nour - they are the "resistance."
- Stephen Vincent
(Journalist murdered by terrorists, 8/2/2005)
Anonymous
August 11, 2005 9:43:58 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Giftzwerg schreef:

> All I'm saying here is that wargames are dependant on real, unscripted
> AI in a way that very few other genres are.

I'd go even a step further and claim that no other genres even have an
AI as advanced as seen in some modern wargames.

Going on a tangent here ...

But as clever as they seem, they are still rule-based - a complex
script, but a script nevertheless. They work within the concept of the
game and even there they still have their limitations simply because
they can't rewrite part of the script themselves.

Enter neural networks - they were all the rage in the early nineties,
but couldn't deliver on the promises they made so are by now almost
completely forgotten. I dabbled with neural network tools and found
them amazingly good at simple and single tasks, but unsuited for the
complexity of let's say a wargame.

Simply put a neural network is a black box function that you train by
providing both input-data and output-data. After the training period is
over you can give it a set of new input-data and receive an answer. The
beauty of the system is that you never had to make the rules. The black
box itself has made them for you and continues to do so and adapts as
long as feedback is given on the output-data (good or bad).

That was over 10 years ago.

I'm also a bit of a FPS fan so I fully expected some adaptive learning
techniques to have entered the realm of the 'bots by now. But it
hasn't.

After a while you "see" the attack patterns of the bots and adapt your
technique after which they become easy prey as they *never* change
their pattern, no matter how many times you kill 'm. Just a little
backward propagation would allow the 'bots to change their pattern once
the negative output results started to kick in.

Short : some neural network techniques would really enhance FPS games

Curving back to wargames ...

Could neural network techniques be used in wargames ? Let me narrow
this question to : Could AI adaptation to the player be used in
wargames ?

Let's say that in a strategic level ETO game the Axis player has this
tendency to perform a Sea-Lion. After being successfull in doing this a
couple of times the player will notice that the Allied AI player
augments his defence of the island maybe resulting in Egypt being
weakened which will create new opportunities for the Axis there - and
so on. This behaviour is no different than when 2 human players play
the same game for a great number of times.

This surely enhances the replayability of a game, so what's the catch ?


Well, as the game will adapt to your style of play, you won't win as
much. Let's face it : we don't mind getting beaten by a machine -
occasionally. The object is to win, we feel better when we win. Winning
less could get propagated back to you not liking the game as much as
you used to. And that's bad for a developer working on his next game.

Another thing is that the game-experience will not be universal, forget
about mods and user-created scenarios as they will behave different
from one computer to the next. This could be compensated for by having
a repository of AI's.

Any remarks and observations highly appreciated

Greetz,

Eddy Sterckx
Anonymous
August 11, 2005 11:34:28 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

I know my interface programmer will have a different view as his
partner is just starting phd on neural networks for FPS games ( you see
your prayers are being ansered ), however I'm very skeptical of the
efficacy of neural networks as a practical decision making tool. I
remember back in 1992, the Australian DoD had set up this think tank
group to develop AI systems for modelling military simulations. They
were very keen on using neural networks. I was doing a contract for
them at the time and they showed me what they were doing. They poured
heaps of money into it but at the end of the day had nothing that
worked.

In the meantime we plodded on using a common sense approach to
designing AI, primarily a rule based system. It might not have had the
"new tech" glamour but it powers the most sophisticated AI engine
available today. :) )
Anonymous
August 11, 2005 3:33:48 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> be*. But AI pilots have a really, really, really long way to go before
> anyone - the public *or the guys who own that $300,000,000 airplane* -
> is going to entrust them with sole responsibility for the flight.

Yeah, sorry, I wasn't debating your comments on wargame AI, I was merely
chucking in a tenuously related titbit that I thought might be amusing.

Interestingly, however, I've recently been playing the Warhammer 40k-based
RTS Dawn of War. Graphically, this is a really beautiful game. However,
the fact that it looks pretty is not all it has going for it.

The games major resource is acquired by taking and holding various Strategic
Points and Critical Locations around the battlefield. Consequently, as
RTS's go, the game is almost completely about strategy and fighting, rather
than resource gathering. I'm not sure if this counts as an evolution in RTS
design but it does show there is more than one way of approaching RTS
development and that RTS games are by no means "developed out".

More importantly, however, while the game's AI straight out of the box is
pretty pathetic, it is highly moddable and a number of enthusiasts have
taken up the challenge. The AI mod for the game that they have produced is
very good and produces an computer opponent that behaves very much like a
competent human player. Combined with a number of options that enable the
gamer to tweak the difficulty of the game to suit their own level of
competence, the game is much more than a click-fest, though obviously, at
the higher levels of on-line play, the ability to use the interface quickly
and efficiently is essential.

My point is that AI development, while clearly one of the most challenging
aspects of game design, is not a reason for wargame mediocrity. RTS games
also have AI issues. I would suggest that, as many have stated, the niche
nature of the serious wargame market is probably the major limitation.
Beautiful but expensive-to-produce user interfaces are simply not
cost-effective in serious wargames. With the right difficulty settings, the
real-time play of Dawn of War is for me as challenging but also as
easy-to-use as RDOA/HTTR. Dawn of War sales, however, are in the tens or
hundreds of thousands. It's a no brainer to work out which game will better
reward more development effort.

Cheers,

Andy
Anonymous
August 11, 2005 3:33:49 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

In article <dde2t3$7s6$1@lust.ihug.co.nz>, andybrn@somewhere.in.nz
says...

> My point is that AI development, while clearly one of the most challenging
> aspects of game design, is not a reason for wargame mediocrity.

I would rephrase that as, "It's not *necessarily* a reason for wargame
mediocrity." Clearly there are some wargames that really shine in the
AI area, and some which have paid little attention to AI.

All I'm saying here is that wargames are dependant on real, unscripted
AI in a way that very few other genres are.

> RTS games
> also have AI issues.

True, but they camouflage this by putting the player under as much time-
pressure as possible. It's as if a chess program which wasn't nearly as
smart as CHESSMASTER 9000 was programmed to only play *speed chess*.

Other genres, ditto. FPS games feature enemies so woodenheaded that one
enemy stormtrooper fails to register his counterpart being shot down
with a blaster five feet away - but that's OK, because any sort of
realism would mean our hero would get exactly five feet. Flight
simulators portray AI enemy pilots as going through an almost random
series of pre-plotted air combat moves - but that's OK, because real
pilots use the same maneuvers, and it looks pretty good. Empire-
building games generate AI opponents who can't continue a strategy from
one turn to the next - but they handicap the AI with bulging coffers of
money and other similar advantages.

Only a wargame needs a strong AI to play against - and the more advanced
wargames (HTTR/CM) are PCT or WEGO and depend on AI routines to command
your *own* troops at times, so their partisans can't even skank out of
AI development by hinting that "PBEM" is the preferred mode of play.

> I would suggest that, as many have stated, the niche
> nature of the serious wargame market is probably the major limitation.

Sure. More people can always be found to play an insipid game that
mimics some slightly-less-tedious-than-their-own version of reality like
THE SIMS than will want to command a historically-correct campaign to
drive Rommel out of North Africa.

Heck, more people like to play Whack-A-Mole, see the "Giant Rats," and
ride some silly rollercoaster at a lame state fair than visit the Patton
Museum of Cavalry and Armor at Fort Knox.

But ultimately ... so what?

--
Giftzwerg
***
"The most despicable misuse of terminology, however, occurs when
Leftists call the Saddamites and foreign jihadists 'the resistance.'
What an example of moral inversion! For the fact is, paramilitary death
squads are attacking the Iraqi people. And those who oppose the killers
- the Iraqi police and National Guardsmen, members of the Allawi
government, people like Nour - they are the "resistance."
- Stephen Vincent
(Journalist murdered by terrorists, 8/2/2005)
Anonymous
August 11, 2005 8:42:16 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

"Arjuna" <dave@panthergames.com> wrote in news:1123770868.161454.162440
@z14g2000cwz.googlegroups.com:

> I know my interface programmer will have a different view as his
> partner is just starting phd on neural networks for FPS games ( you
see
> your prayers are being ansered )

If they need some overpriced consultancy on this, they know where to ask
:) 

>, however I'm very skeptical of the
> efficacy of neural networks as a practical decision making tool.

Neural networks were over-hyped back then and marketed as a tool that
could fix everything. The backlash when that bubble bursted meant that
from then on it was pretty much ignored from then on to this day. Given
that both the available processing power has significantly increased and
that the expensive lessson of what it can and can't do has been learnt
it might make a come-back in very specific areas.

In the specific case of the Airborne Assault engine I will type-up a
feature request/proposal/scheme after COTA is out the door. It will
probably end up in the "pile of perpetual ignore" as most of my other
wild ideas, but at least it will allow the other beta-bunnies to have a
bit of fun by poking holes in my theories :) 

> They poured
> heaps of money into it but at the end of the day had nothing that
> worked.

Well, my earliest experience was a bit different. We had this general
purpose Neural Network program which we were supposed to evaluate for a
major bank. This bank had a cafetaria where every day there were 3
different lunches you could order. What we did was feed the NN program
with the raw data of each menu (input) and which of the 3 I would pick
for that day (feedback). It literally went like :

menu 1 : fries, steak, rocquefort sauce
menu 2 : cod, mashed potatoes, bearnaise sauce
menu 3 : noodles, shrimps, tomato sauce
pick : menu 1

After training the NN like this for a couple of weeks it could pick out
the menu I would choose 90% of the time. Which was at the same time
pretty amazing and not good enough (for business purposes)

The thing is the NN would pick the correct menu even if it was the first
time it appeared. i.e. it could work outside the bounderies of a rule-
based system.

Even then I wasn't stupid enough to think it would change the world as I
was well aware of the inherent limitations (very narrow applicability,
needing lots of training = time = money, accuracy) but nevertheless it's
something that I think is being unjustly ignored and ridiculed.

Then again, I might be completely wrong about this.

Greetz,

Eddy Sterckx

--
"Ceterum censeo Belgicam delendam esse."
(Cato, 'Pro Gerolphe')
Anonymous
August 11, 2005 9:09:16 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

>If they need some overpriced consultancy on this, they know where to ask
:) 

Eddy, it's for a PhD not Defence. ;) 
Anonymous
August 12, 2005 5:52:48 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> Spending another 200K on graphics will not triple their market ...
> There are worse niche markets than targeting a group of middle-aged
> computer-owners with more money then time.

Out frugal, out last, or out live the competition. Works for me. :) 

Best regards, Major H.
tacops@mac.com
http://www.battlefront.com/
Anonymous
August 13, 2005 7:21:10 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Andy,

And that's what's wrong with puting out a demo. Unless it can easily,
readily and comrehensively introduce the game without the need to read
a manual it's counterproductive as far asa marketing tool goes.
Anonymous
August 13, 2005 3:20:07 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> Mass commercial appeal is fine. Great graphics are ... well, great.
> Slick interfaces make everyone happy. But often design choices have to
> be made that present a clear choice to the wargame designer; do I go for
> the rigorous historical accuracy that will impress the "grogs," or do I
> incorporate an abstraction that will please the broader market?

But sooner or later, we're going to be presented with two wargames, equally
rigorous in thier historical accuracy, one of which has a prettier interface
than the other. Because we're enthusiasts, we'll probably buy both of them,
but can we speak for everybody?

Interestingly, I downloaded the GGWaW demo the other day and I notice that a
bit of effort seems to have been put into its interface. It's not a work of
art but it's certainly more than just "functional", and those video clips
aren't really needed for "gameplay", are they? Have 2by3 read the writing
on the wall?

Cheers,

Andy
Anonymous
August 13, 2005 3:20:08 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

In article <ddjar3$idb$1@lust.ihug.co.nz>, andybrn@somewhere.in.nz
says...
> > Mass commercial appeal is fine. Great graphics are ... well, great.
> > Slick interfaces make everyone happy. But often design choices have to
> > be made that present a clear choice to the wargame designer; do I go for
> > the rigorous historical accuracy that will impress the "grogs," or do I
> > incorporate an abstraction that will please the broader market?
>
> But sooner or later, we're going to be presented with two wargames, equally
> rigorous in thier historical accuracy, one of which has a prettier interface
> than the other. Because we're enthusiasts, we'll probably buy both of them,
> but can we speak for everybody?

Oh, I'll happily agree that *all other things being equal*, I'd pick the
wargame with the slicker interface and better graphics.

The trouble is, this is like saying, "all other things being equal," I'd
pick the car with the massive engine and the all-power package. This is
fine - except that "all things" are rarely "equal."

> Interestingly, I downloaded the GGWaW demo the other day and I notice that a
> bit of effort seems to have been put into its interface. It's not a work of
> art but it's certainly more than just "functional", and those video clips
> aren't really needed for "gameplay", are they? Have 2by3 read the writing
> on the wall?

I gave up on GGWAW. Even when one turns off all that useless eye-candy,
it still compares poorly to STRATEGIC COMMAND. Even 3RPC is a better
option, IMO.

--
Giftzwerg
***
"The most despicable misuse of terminology, however, occurs when
Leftists call the Saddamites and foreign jihadists 'the resistance.'
What an example of moral inversion! For the fact is, paramilitary death
squads are attacking the Iraqi people. And those who oppose the killers
- the Iraqi police and National Guardsmen, members of the Allawi
government, people like Nour - they are the "resistance."
- Stephen Vincent
(Journalist murdered by terrorists, 8/2/2005)
Anonymous
August 13, 2005 7:48:08 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> I gave up on GGWAW. Even when one turns off all that useless eye-candy,
> it still compares poorly to STRATEGIC COMMAND. Even 3RPC is a better
> option, IMO.

I'm not too impressed with it myself, although I've only got the demo to go
on. Probably a good example of how not to do a demo? The tutorials are
fine as far as they go but they leave too much unsaid, especially about
supply which I suspect is a vital component of the game. The manual is
included for me to read but I don't feel like wading through 120-odd pages
just to find out if I like a game.

Andy
Anonymous
August 13, 2005 10:05:13 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Andy,

Fair enough. Remember though that it costs a lot of money to do a good
effective demo.
Anonymous
August 14, 2005 10:35:11 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Well Any that's very nice to hear. Thank you.

For what it's worth we embedded a survey in most of the demos that went
out and while you only ever get a small percentage of downloaders
actually responding the vast majority of those who did were polarised
into two camps - those that thought it way too complex and those that
thought it way too simple. lIRC only 2% bothered to play the tutorial
and almost noone read the pdf manual that came with it.
Anonymous
August 14, 2005 10:35:16 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Well Andy that's very nice to hear. Thank you.

For what it's worth we embedded a survey in most of the demos that went
out and while you only ever get a small percentage of downloaders
actually responding the vast majority of those who did were polarised
into two camps - those that thought it way too complex and those that
thought it way too simple. lIRC only 2% bothered to play the tutorial
and almost noone read the pdf manual that came with it.
Anonymous
August 14, 2005 10:35:18 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Well Andy that's very nice to hear. Thank you.

For what it's worth we embedded a survey in most of the demos that went
out and while you only ever get a small percentage of downloaders
actually responding the vast majority of those who did were polarised
into two camps - those that thought it way too complex and those that
thought it way too simple. lIRC only 2% bothered to play the tutorial
and almost noone read the pdf manual that came with it.
Anonymous
August 14, 2005 3:44:59 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> And that's what's wrong with puting out a demo. Unless it can easily,
> readily and comrehensively introduce the game without the need to read
> a manual it's counterproductive as far asa marketing tool goes.

I felt the GGWaW demo did a good job of showing me how to do things but it
didn't explain to me why I'd want to do them. The cause-and-effect between
building X, moving Y and winning the game was NOT clearly demonstrated.

OTOH, there was no way I would have bought the game without getting a look
at it first so this particular demo has merely confirmed my decision not to
buy. It has not cost 2by3 a sale. I feel better about the company because
they produced it, I just wish they could have done it better.

Andy
Anonymous
August 14, 2005 8:02:47 PM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Oops sorry for the treble post. :-(
Anonymous
August 15, 2005 3:58:12 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

> Fair enough. Remember though that it costs a lot of money to do a good
> effective demo.

For what it's worth, I would never have bought RDOA if it wasn't for the
demo. It was Just Another Arnhem Game as far as I was concerned. Folks
were commenting favorably on it but that wasn't enough to tempt me.

It wasn't until I downloaded the demo (let's see what all the fuss is about)
that I realised how innovative it was. I ordered it the next day and
remember being very impatient waiting for it to arrive by post.

Andy
Anonymous
August 23, 2005 6:35:49 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Gregor Whiley schreef:
> but it's not an area that is
> susceptible to overnight breakthroughs, or panaceas like neural
> networks.

Agreed that "Combat AI" (frontline detection, path-finding, combat odds
optimization, ...) have nothing to gain from NN not only because
existing techniques like influence mapping, A* shortest path routines
and simple brute force calculation are already very good at this but
also because NN is a tool unfit for this type of AI.

Moreover - the very best of the engines (your's included) have already
licked the "Combat AI" to a degree that the game AI reacts like a
competent player which is all you can really ask for.

You've correctly identified the "strategic" AI, the decision maker as
the area in which a lot of progress can still be made and while I'm not
saying that NN is the deus ex machina that will solve all your problems
there, I believe certain NN techniques that could be beneficial are
overlooked because of the bad rep NN has gotten with developers.

Let's use an example here : let's say in a particular scenario your
strategic level AI has to make a choice regarding the defence of a
major victory location and the AI has to make the choice of either a
heavy forward defense or a light forward screen with all the big units
hogging the victory location. Given the current level of strategic
level AI this decision is taken randomly. Now, a human player will have
a certain attack style - let's say that some players have a methodical
broad front strategy, others opt for a rapier like thrust.

That gives the following matrix :

Forward defense + thrust => human win
Forward defense + broad attack => ai win
Victory location defense + thrust => ai win
Victory location defense + broad attack => human win

Now, as I said : humans tend to have a certain playing style, in this
example let's say it's the "thrust" so with a randomly picked defensive
strategy the AI will lose about half of the time. Now let's introduce a
little NN with a bit of backward propagation. When the scenario is
played for the first couple of times the defensive strategy will be
just the same as a randomly picked defensive strategy but then the
learning process starts to kick in and the NN will learn that "victory
location defense" has a higher probability of winning the game for it -
even without it needing to formaly identify the player's style as
"thrust" (this bit is important), because for the NN the input is
simply "picked victory location defense => win", "picked forward
defense => lose". The weighing process in a NN will make it go for the
strategy to best counter that particular human player giving him a
better game. The data-file to feed such a basic NN is pretty simple.

I've identified some obvious problems with this approach (scenario's
not playing universal anymore, player maybe not liking the AI getting
smarter/adaptive, ...) and I'm certainly not advocating to make
everything NN, but I will try to push these backward propagation
techniques where appropriate.

Now a cynic might suggest I use a bit of this backward propagation
techniques on my own ideas. Here's me thinking some NN techniques might
be usefull and both bleeding edge AI engine designers (SSG and Panther)
tell me it's a dumb idea and I still keep pushing it ... :) 

Greetz,

Eddy Sterckx
Anonymous
August 23, 2005 8:50:42 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

eddysterckx@hotmail.com wrote:

> Now a cynic might suggest I use a bit of this backward propagation
> techniques on my own ideas. Here's me thinking some NN techniques might
> be usefull and both bleeding edge AI engine designers (SSG and Panther)
> tell me it's a dumb idea and I still keep pushing it ... :) 
>
> Greetz,
>
> Eddy Sterckx

LOL - you said it.

One of the more obvious difficulties with your proposal is that there
area myriad of factors that could have brought about success or
failure. Further there is usually a pretty broad spectrum between
absolute success and absolute failure. Moreover, unless you can
adequately identify specific sitations, specific options/strategies and
the specific outcomes then you cannot apply this to a command system
game like HTTR, other than at the top or side level.
Anonymous
August 23, 2005 9:16:31 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

Arjuna schreef:

> One of the more obvious difficulties with your proposal is that there
> area myriad of factors that could have brought about success or
> failure. Further there is usually a pretty broad spectrum between
> absolute success and absolute failure.

I used an extremely simple example to make my point, the beauty of a NN
system is that it really doesn't care if it has to work with ranges or
discrete values. Moreover it doesn't care about the reason for success
or failure or even about the logic behind it. Quantify an AI decision
and note the victory points at scenario end is all it takes grosso
modo.

> Moreover, unless you can
> adequately identify specific sitations, specific options/strategies and
> the specific outcomes then you cannot apply this to a command system
> game like HTTR, other than at the top or side level.

You fully well know that no amount of reasonable arguments will stop me
from typing up a Panther engine specific proposal / feature request
once COTA is out the door. It's Neural Networks that act reasonable,
not me :) 

Greetz,

Eddy Sterckx
Anonymous
August 23, 2005 9:20:23 AM

Archived from groups: comp.sys.ibm.pc.games.war-historical (More info?)

On 8 Aug 2005 00:18:36 -0700, "eddysterckx@hotmail.com"
<eddysterckx@hotmail.com> wrote:

>Hi,
>
>The StrategyZone's Don Maddox interviews Dave O'Connor, John Tiller,
>Gregor Whiley, Mark H. Walker and Patrick Proctor on the current state
>of wargaming and what they think the future will bring.
>

I just thought I'd say something about AI routines for wargames, in
light of the preceding discussion.

There are parts of our AI that benefit from a brute force approach,
and thus from increases in computing power. The most obvious example
is the AI combat routines/Combat Advisor which are virtually
identical. As people who have played the game will know, the Combat
Advisor shows the best possible odds for a combat, and shows which
units need to move to where to achieve these odds. It is highly
accurate, though the very best human players can sometimes finesse
slightly better results than the Combat Advisor can achieve. Note that
the Combat Advisor won't tell you if it's a good idea to proceed with
the combat, just the maximum achievable odds.

However, it's certainly smarter than me, and is used by all players,
even the very best, so I know that my claims of accuracy are
reasonable. To get these results, the AI notionally moves every unit
that could potentially take part in the combat until it detects the
best possible odds (or a combat with maximum odds). This takes a fair
amount of processing power to achieve in an acceptable time, and would
not have been possible on older machines.

The AI also relies heavily on the AI movement routines to tell it
which units can actually reach the combat, and these likewise use a
brute force approach to guarantee the lowest possible cost for any
movement path, something that also benefits human players.

As a result, I think that the AI routines for combat that we have in
the Decisive Battles series is about as good as you can get,
considered on a purely tactical level of maximising combat odds.

What's much harder, is the decision, alluded to above, as to whether
it is a wise decision to make the attack. As wargamers, you will know
the large number of factors that go into making that decision. The
computer that can make a wise decision in a wargame has yet to be
built. The best computers can do is make the same decision that a
human would, but they are always going to make those decisions for
different reasons and in different ways.

It is at the more strategic level of AI where progress is not really
dependent on processing power. We feel that we have made progress in
this area, and that the AI in say Battles in Italy is better than it
was in the original Ardennes Offensive, but it's not an area that is
susceptible to overnight breakthroughs, or panaceas like neural
networks.

Gregor


Gregor Whiley
Vice President, Strategic Studies Group
http://www.ssg.com.au
!