The Future Of Complex Wargames Looks Bleak

[Missouri_Rebel]

ORIGINAL: Arjuna
Well you do not what you are talking about I'm afraid to say. Our Command ops is not scripted and I think most would agree that it fills the buill as being a complex game.

O.K. Moderators. Let's keep the discussion civil. [:D]

[Arjuna]

Yes it does depend on how you define scripting. I think most would agree that scripting involves executing a series of actions in a sequence. In most games scripts are precanned event sequences. There maybe some conditional logic that at a certain point executes one sequence or another but ultimately they are still precanned sequences. In most games the AI is not capable of developing a plan on the fly itself. It merely decides on which of the pre-canned plans it wishes to implement at that point in time.

In most cases the highest entity (ie the side) will have situational awareness in that it can interrogate the virtual world but subordinate entities (ie the units) only have very limited situational awareness and can at best react to immediate events - eg they can take cover iof fired upon. In many games the subordinates have no flexibility to determine their own course of action. They are in effect dumb entities slaved to the side's script. In a few games some of the entities may have their own scripts which form subsets of the side's script or they may be independent from the side's script but they are precanned scripts nonetheless.

In Command Ops each unit is situationally aware and can develop its own course of action based on the virtual world around it. The only "guides" we use are the side objectives. Note that these are generic tasks like Secure X from time 1 to time 2. There is no specific course of action given as to how they should do this or with what forces. There are victory points assigned to the side objectives but these act only as a guide to the AI in determining how it should allocate its forces to each of the objectives. In fact the AI can reprioritise based on developments. For instance, it can increase the priority of one objective if it is threatened by the enemy and decrease another if it is not. In fact it can abandon objectives if it thinks it has no chance of achieving them or it can sequence their achievement if it thinks its best to take certain objectives one at a time rather than concurrently.

But the real distinguishing feature here is that in a generic AI each entity can reassess and develop a new plan if it thinks that it will be better to do so. A game that relies on a scripted AI doesn't model the full decision loop. It executes a plan someone else has provided. It may be able to react but it can't reassess dynamically only if certain scenario specific events or conditions change. So if the battle unfolds different to the way the designer has scripted things it cannot take advantage nor minimise adverse effects. It is a captive of the original design and concept of operations.

[Missouri_Rebel]

ORIGINAL: Arjuna

Yes it does depend on how you define scripting. I think most would agree that scripting involves executing a series of actions in a sequence. In most games scripts are precanned event sequences. There maybe some conditional logic that at a certain point executes one sequence or another but ultimately they are still precanned sequences. In most games the AI is not capable of developing a plan on the fly itself. It merely decides on which of the pre-canned plans it wishes to implement at that point in time.

In most cases the highest entity (ie the side) will have situational awareness in that it can interrogate the virtual world but subordinate entities (ie the units) only have very limited situational awareness and can at best react to immediate events - eg they can take cover iof fired upon. In many games the subordinates have no flexibility to determine their own course of action. They are in effect dumb entities slaved to the side's script. In a few games some of the entities may have their own scripts which form subsets of the side's script or they may be independent from the side's script but they are precanned scripts nonetheless.

In Command Ops each unit is situationally aware and can develop its own course of action based on the virtual world around it. The only "guides" we use are the side objectives. Note that these are generic tasks like Secure X from time 1 to time 2. There is no specific course of action given as to how they should do this or with what forces. There are victory points assigned to the side objectives but these act only as a guide to the AI in determining how it should allocate its forces to each of the objectives. In fact the AI can reprioritise based on developments. For instance, it can increase the priority of one objective if it is threatened by the enemy and decrease another if it is not. In fact it can abandon objectives if it thinks it has no chance of achieving them or it can sequence their achievement if it thinks its best to take certain objectives one at a time rather than concurrently.

But the real distinguishing feature here is that in a generic AI each entity can reassess and develop a new plan if it thinks that it will be better to do so. A game that relies on a scripted AI doesn't model the full decision loop. It executes a plan someone else has provided. It may be able to react but it can't reassess dynamically only if certain scenario specific events or conditions change. So if the battle unfolds different to the way the designer has scripted things it cannot take advantage nor minimise adverse effects. It is a captive of the original design and concept of operations.

And that unit does what with the information? Is it passed up the chain of command in Command Ops? Is it instantly known to all the forces and the AI in general?

My guess is that even if it does react to the changing situation that it is still defined by the choice of 'scripts' it has to respond to that specific parameter and that becomes the most important aspect.

Am I to assume that those reactions are different depending on the level of the affected unit? Would that then need a whole new set of 'scripts' with the information possibly going down the chain of command?

Hard to wrap my head around it.

Mo reb

[Erik Rutins]

ORIGINAL: Missouri_Rebel
My fear is that along with everything else it seems that wargames will become increasingly more simple and lack depth.
Publishers and game makers don't seem to be making the 'epic' wargames of yesteryear with a few exceptions. And of those that are released it appears that they fail in terms of wide appeal for a few reasons.

Well, I have to say from where I'm sitting it doesn't look like that. We have just as many, if not more, complex wargames in the works and planned for the future now than at any time in the past. It also happens that we have more games that are also less complex planned. I think you need to cater to both sides of the market, but making more non-complex games is not the same as making fewer complex games. I can't speak for other publishers, but one of the reasons we created Matrix Games was specifically to serve the community that enjoys complex and realistic strategy games.

In short, as long as there are still gamers interested in playing these games, we plan to be making them.

Regards,

- Erik

[Erik Rutins]

I also think it's quite possible to make a good non-scripted complex game AI and that's been proven by Command Ops and several other titles as well. However, many titles do rely on some limited or partial scripting to help the AI in certain situations. That's what Iain was referencing. For example, in War in the East the AI uses a script for the opening moves, but for the other 99% of the turns it's on its own, nevertheless this would count as "partially scripted". Scripting in an of itself is not a bad word where AI is concerned - it's just another tool and another form of programming. Scripts can range from "here's the entire turn, just do this" to small expert modules that the AI has to know how to assemble to create a larger strategy. While my experience with AI is not as a developer, I've worked with many AIs from the design side and scripting is another tool along with many possible AI implementations that may fit one game better than another.

Regards,

- Erik

[Erik Rutins]

We did have a relationship with VASSAL, but that was years ago and it is no longer the case.

Regards,

- Erik

[Missouri_Rebel]

Well, I have to say from where I'm sitting it doesn't look like that. We have just as many, if not more, complex wargames in the works and planned for the future now than at any time in the past. It also happens that we have more games that are also less complex planned. I think you need to cater to both sides of the market, but making more non-complex games is not the same as making fewer complex games. I can't speak for other publishers, but one of the reasons we created Matrix Games was specifically to serve the community that enjoys complex and realistic strategy games.

In short, as long as there are still gamers interested in playing these games, we plan to be making them.

Regards,

- Erik

That is good to know. I am not espousing some kind of wargame snobbery either. I buy and play many of the less complex games because, quite frankly, I'm not always up to racking my brain out with every important move.

I just want to know that there is a future with games of considerable depth and I'm glad to hear that from where you're sitting things look positive. Hopefully that trend will continue.

mo reb

[ComradeP]

Although the Command Ops AI is good, problems with units halting or companies attacking in an illogical manner, which were fixed and/or introduced and fixed with the recent patch do cripple the system somewhat. If your system depends on a good AI, it collapses quickly if the AI doesn't work as intended.

The Command Ops AI would presumably also work less well on a strategic scale (which is where most AI's fail, because they are incapable of making good long term plans). Battles are fairly short, distances are not that significant. It's also different in the sense that the destruction of units/causing casualties to the enemy often doesn't give a lot of points and that objectives are captured in a somewhat abstract manner which took me some serious getting used to. It's unique amongst wargames, but the system has its own weaknesses which would make it function less well on a strategic scale in its current shape.

As I only own BftB, I'm also not entirely sure how the system would work for more evenly matched opponents.

[Arjuna]

Well smart complex wargames require a significant investment of resources to make them "smart". That cost a lot and for that cost to be justified the games have to sell more that a lightweight game or they have to be sold at a higher price. Any serious developer venturing into this business would probably conclude that given the past reaction here to any increase in price that its simply not viable economically to develop "smart" complex wargames. They would only do so if they were passionate about the subject and put that above making money. Yes there are a few of us fools doing so but don't be surprised that there isn't a rush of developers working on complex wargames.

[Erik Rutins]

Just in case any of you do not already own and play Command Ops, if you have any interest in the subject matter please get a copy. It is IMHO an indispensible title for any serious wargamer.

Regards,

- Erik

[Arjuna]

ORIGINAL: ComradeP

Although the Command Ops AI is good, problems with units halting or companies attacking in an illogical manner, which were fixed and/or introduced and fixed with the recent patch do cripple the system somewhat. If your system depends on a good AI, it collapses quickly if the AI doesn't work as intended.

Agreed, which is why we put in the massive effort to iron out those shortcomings.

The Command Ops AI would presumably also work less well on a strategic scale (which is where most AI's fail, because they are incapable of making good long term plans). Battles are fairly short, distances are not that significant.

Actually the Command Ops AI is designed to scale. The notion of distances not being great is misleading. What really matters is how many grids you have in your map area. Currently our recommended max map size is 5000 x 4000 grids. If each grid was 2kms instead of 100m then you're looking at being able to model a map area 10,000 kms x 8,000 kms. The route fining would be just as good. Similarly, if you rescale the time interval from representing one minute to 2 hours then a ten day battle at the operational level becomes 1200 days.

It's also different in the sense that the destruction of units/causing casualties to the enemy often doesn't give a lot of points

That is just a scenario designer setting. In the ScenMaker you can set the VPs awarded to enemy destruction to whatever you like.

and that objectives are captured in a somewhat abstract manner which took me some serious getting used to.

What do you mean by "abstract manner"? To achieve an objective depends on its type. For a Secure objective all you have to do is have frinedly forces in the objective perimeter that outnumber the enemy forces 10:1. That's pretty simple really. For a Defend objective you just have some forces within the perimeter. That's even simpler.

It's unique amongst wargames, but the system has its own weaknesses which would make it function less well on a strategic scale in its current shape.

I admit that developing a strategic version would take time. There are features that would have to be redesigned or added like political realtionships. I recently went through this exercise - see the discussion on War-Histroical about Funding Trial of Strength here:
https://groups.google.com/forum/?hl=en& ... jynjqtgDnU
So basically I think you are wrong on this point. But I suppose that until we actually do a strategic level version there is no way of proving this one way or the other.

[Arjuna]

ORIGINAL: Erik Rutins

Scripting in an of itself is not a bad word where AI is concerned - it's just another tool and another form of programming. Scripts can range from "here's the entire turn, just do this" to small expert modules that the AI has to know how to assemble to create a larger strategy. While my experience with AI is not as a developer, I've worked with many AIs from the design side and scripting is another tool along with many possible AI implementations that may fit one game better than another.

Regards,

- Erik

I agree Erik. Scripts are but one tool and they do have their uses. The military use them to control the nsequence of events for training. They want things to be predictable so they can cover all their training objectives. They can be very handy for modeling historical events or opening moves that yopu want to have in place. In fact such a use has been on our wish list for some time to implement into Command Ops.

[Perturabo]

Another thing that I like about Command Ops is that it also sometimes takes place in an exotic, rarely explored theatre. I'm looking forward to finally being able to visit France in a real time simulation game.

[Saint Ruth]

ORIGINAL: catwhoorg

Commander - the Great War is a good example of a challenging AI in a recent game

Its not as inventive as a human, but I certainly provides an interesting enough Single player game

It is a good game against an AI. Perhaps though it was designed with the AI in mind? Some of the rules benefit an AI (nothing wrong with that if the rules apply to the human too). E.g. every city is a supply source. oh, that's a "Command Europe at war" rule, not sure if same for that game. So the Russians can defend every city and build fronts from a city outwards so surrounding is only a problem if there's no city nearby.
Nothing wrong with that of course. Perhaps it's the AI way to go.

[Saint Ruth]

ORIGINAL: sterckxe
It's not mythical at all - I've played the beta - my review is up here :

tm.asp?m=891383&mpage=1&key=&#

Very interesting...always wondered what happened to that game. Still, maybe there'll be a game like that for my retirement!

[ComradeP]

That is just a scenario designer setting. In the ScenMaker you can set the VPs awarded to enemy destruction to whatever you like.

Perhaps, but the game still has the usual problem of operational level wargames that the destruction of an enemy force is largely irrelevant because most of the points come from capturing objectives. That's why I don't think it would scale well in its current shape to a more strategic level: if battles last a month or two, in good weather, it matters if the defender loses most of a division in a sector in the opening days. Currently, I can wreck the Allied force and still lose with relative ease due to not holding the final objectives for long enough.

That remains one of the main differences between operational and strategic wargames: in a strategic level wargame, serious losses are much more of a problem, because you have to deal with the initial losses for the entire battle.

The current artillery system, of being able to throw the kitchen sink at any target of choice, would also cause problems on a strategic scale.

What do you mean by "abstract manner"? To achieve an objective depends on its type. For a Secure objective all you have to do is have frinedly forces in the objective perimeter that outnumber the enemy forces 10:1. That's pretty simple really. For a Defend objective you just have some forces within the perimeter. That's even simpler.

It's abstract in the sense that a point counter starts when you first take hold of an objective, which determines the points you get for an objective (independent of friendly unit strength it seems, which is something I don't entirely like as 1 platoon theoretically captures an objective as quickly as an entire division as long as there's no enemy around). That is more abstract than the usual "you get the points if you capture/hold it" approach.

[KG Erwin]

Frankly, even though I've been playing wargames for over 40 years, overly complex and "monster" games are not my cup of tea. The biggest board wargame I ever bought was SPI's War in the East, which was designed by Jim Dunnigan, and while I set it up a couple of times, the sheer size of it was overwhelming. I prefer PC games these days, and the most complex one I currently play is BOA 2.

That being said, while I'm an Eastern Front fan, I hold no desire to play Gary Grigsby's WiTE. I say this simply because I want to have time for my real life. If I could afford to be retired for my day job, I might consider it.

[berto]

ORIGINAL: KG Erwin

That being said, while I'm an Eastern Front fan, I hold no desire to play Gary Grigsby's WiTE. I say this simply because I want to have time for my real life. If I could afford to be retired for my day job, I might consider it.

+1

Yes, playing a monster game (like WITP:AE; I only play the scenarios) tends to crowd out all other gaming, and much of Real Life besides.

I don't like beer & pretzel games. So I take the middle ground.

[Ranger33]

Same here, I much prefer the idea of monster wargames to actually playing them [:)]

Right now I'm in the end stages of pharmacy school, so I don't have time for much gaming at all, certainly not overseeing the entire Eastern Front. I do own WiTE and found it much easier to get into than I expected (as far as the tutorial and smaller scenarios) and very enjoyable to play, but I simply don't have the time to have a go at the big show.

[wodin]

Actually many scenarios have an amount of points for destroyed enemy..it's down to the scen maker...quite easy for you to get what you want..give 100 points for enemy destroyed.

ORIGINAL: ComradeP

That is just a scenario designer setting. In the ScenMaker you can set the VPs awarded to enemy destruction to whatever you like.

Perhaps, but the game still has the usual problem of operational level wargames that the destruction of an enemy force is largely irrelevant because most of the points come from capturing objectives. That's why I don't think it would scale well in its current shape to a more strategic level: if battles last a month or two, in good weather, it matters if the defender loses most of a division in a sector in the opening days. Currently, I can wreck the Allied force and still lose with relative ease due to not holding the final objectives for long enough.

That remains one of the main differences between operational and strategic wargames: in a strategic level wargame, serious losses are much more of a problem, because you have to deal with the initial losses for the entire battle.

The current artillery system, of being able to throw the kitchen sink at any target of choice, would also cause problems on a strategic scale.

What do you mean by "abstract manner"? To achieve an objective depends on its type. For a Secure objective all you have to do is have frinedly forces in the objective perimeter that outnumber the enemy forces 10:1. That's pretty simple really. For a Defend objective you just have some forces within the perimeter. That's even simpler.

It's abstract in the sense that a point counter starts when you first take hold of an objective, which determines the points you get for an objective (independent of friendly unit strength it seems, which is something I don't entirely like as 1 platoon theoretically captures an objective as quickly as an entire division as long as there's no enemy around). That is more abstract than the usual "you get the points if you capture/hold it" approach.