[tactical combat] Terrain influence and other tactical suggestions

from Elemental Forums

I strongly aggree with Citywolfdreams' proposals in the 'War of magic' forum. Tactical battles are such an important and repetitive thing in FE, they should be more interesting! To be honest, I don't own the beta version of the game and only watched "Let's play" videos so far. So, maybe I got a wrong impression of tactical combat and I don't know all details for sure. Therefore, please forgive me, if I talk nonsense at some point.

Anyway, let me share my comments and ideas:

As citywolfdreams proposed, I think tactical combat needs terrain influence. Terrain tiles should be divided into several categories:

- Terrain with no movement penalty, no fighting penalty, no range attack obstacles ("normal", plain grass tiles)

- Terrain with movement penalty, no fighting penalty, no range attack obstacles (rubble, for example)

- Terrain with movement penalty, with fighting penalty, no range attack obstacles (swamps, for example)

- Terrain with movement penalty, with fighting penalty, intermediate range attack obstacles (scrubs, for example)

and

- impassable Terrain with no range attack obstacles (water, for example)

- impassable Terrain with intermediate range attack obstacles (bolders, for example)

- impassable Terrain with full range attack obstacles (trees, for example)

 

Other proposals:

- Impact of height (there are such nice elevations in at the tactical map. USE them!)

- In  addition to that, I think FE needs a (severe!) accuracy reduction for non-magic range attacks based on the distance to the target. (A VERY important point, in my regards - even if ALL OTHER proposals are not taken into account!) This would reduce the "problem" of beeing attacked by enemy archers while moving to a better suited location with the own archers and meele units, too!

- Another way (as additional option) to make it easier to reach final battel formation might be the introduction of a "fast moving target": If an unit moved 3 (maybe 2) or more tiles in it's turn, archers suffer an accuracy penalty.

- If I go boisterously, I could propose a line of sight based on range attacks obstackles. Magic would have no range penalty, but would only be possible, if the target can be seen by the mage (or at least on own unit).

- Finaly, there should be a free attack, when the contrahent is leaving a ZOC (when fleeing or because he wants to atttack another enemy).

 

To be honest, I don't know, how difficult this would be to code. I just know, that - at least in my regards - this would make combat way more interresting.

 

Cheers,

Deggial

 

---

Edit: Why are there parts colored in beige and work as (fake) links? Sorry, this is not by intention...

Edit 2: Well, the joy of the search function... I posted this already, but in the wrong forum. Sorry for the double post, but I really think, this should be in the correct section!

5,477 views 3 replies
Reply #1 from Elemental Forums Top

All of these are good suggestions.  And the rules are relatively simple to implement.  The problem is with designing A.I. that works well with them.

Trust me on this one.  One of our term project's goals in college was a 'simple' tactical game which had nothing but Zone of Control, Defending, Charge, and Opportunity Attacks.  We were a four person team, three of whom were a team that made it twice to the ACM college programming finals. 

We must have put at least at total 400-500 hours into it, and compared to FE's combat, it was trivial. Four combatant types (randomly selected team) three combat stats, one attack type, four rules, and only one victory condition (last man standing)

And to blow me own horn, our team won the class contest, 3:1 in the final, the loss because the program timed out once.  One thing I noticed is that our program was encountering AIs that would try to implement strategies.  Ours was brute force - very optimized, 3-4 turns look ahead, with one goal - max damage out for min damage in.  And it was kicking ass.  But against a human, this approach falls on its face. Once you figure out how it plays, you start offering the AI fast soft targets and it falls into traps or you deny it any action for a few turns and it runs out of look ahead.

Blah, I am babbling. In any case, what I am saying is that the more varied the actions you can take, the harder it is teach the AI to work well with even simple rules. 

You cannot possibly imagine how many real bright people spent '92-'94 trying to make a program that plays Warhammer Fantasy so it's halfway interesting.  It took one week to make the board (the program that handles the rules) but by the time I left, we had nothing that played a good game.  And we are talking 6! turns with a limited set of rules.

Reply #2 from Elemental Forums Top

Thank you for your elaborate answer, Tuidjy.

I don't question your experience and knowledge in programming a tactical combat AI and therefore don't disbelieve in your judgment. Even more, I can not argue at the same level of knowledge, as my own experience is (mostly) that of an user of computer games, not that of a programmer.

All in all, I think, it boils down to the question: "How clever do I want the AI to use the given rule set?"

Personally, I am willing to live with a far from perfect use! I am new to this forum, but if you would know me from the civfanatics (Civilization 5, for those few, who might not know the forum), you would know that I don't have that high expectations of what a computer program is able to achieve in tactical combat at all!  I am pretty aware of the limitations given by a complex rule set and I am fine with the idea of AI bonuses (e.g. in numbers) to compensate the human supremacy. ;)

But I believe, that the tolerance to AS (artificial stupidity) is a personal attitude and I can perfectly understand the wish for equal preconditions for both: AI and human player!

So, let's label my suggestions as "Would be great to have, but they are illusionary." OK, I could live with that. Keeping them in mind as a nice, but unachievable utopia, we could limit our wishes to what is possible and necessary. And here, we come to the main reason for my first post:

When I saw an archer shooting through a forest in a "Let's Play" video, I was shocked! Sorry, but in all willingness for an AI streamlined programming, this is a simple No Go! It was such an immersion breaker that, if this stays in the final game like that, my shortly inflamed love for FE might burn down to ashes instantly. As I said in my initial post: tactical combat is such an important part of the game! Having an at least a halfway decent system, is crucial for the game. At least in my eyes.

Even more, there are games already on the market that can pass at least this task with flying colors - and they are old: Remember "UFO - Enemy unknown" and "Jagged Alliance"? Boy, where these games fun to play! (And there are/will be new incarnations of the game, as we all know.) Both games lived, because of their fun tactical combat. And their battle maps where larger and more complex than that of FE. But I don't even ask for such a decent battle system. Just:

Let's forget about the proposed movement restrictions, if AI can not handle them in a decent way. Let's forget fighting mali (I think they are called "negative bonuses" in English?) on certain tiles, if this is too complicated for the AI to use. But please, PLEASE overwork the archery algorithm! Don't let shoot archers through forests (and other impenetrable objects)! It is not even necessary to teach the AI to make good use of the cover. Just let it use it "by accident". I would be fine with this.

(And, maybe in addition, introduce my proposed range limitations for archers. I think, even *I* could program an algorithm that teaches an archery unit to approach the enemy until a certain efficiency threshold (lets say 75% of their maximal accuracy) is reached. This would make TC more interesting, without rising impassable barriers for the AI.)

-- 

Sorry for being so annoyingly stubborn in this point. But, after all, this is the FE Wishlist thread, isn't it? And this is my MAYOR whish!

SoundPackager - Customize the Windows sound scheme
Reply #3 from Elemental Forums Top

I'm scared to wish for tactical changes.  I've seen someone suggested a feature (I want) of highlighting the 'que' card highlights the unit on the board.  Im not sure if the AI takes into account initiative and battle order or not, but since WE can it would be nice. 

I also cringe a little bit about the 'tactics' in the tactical battle. 

Ranged weapons are as effective at every range.  - this may be a slippery slope if different ranged weapons have different 'ranges' so I understand a simplification but it doesn't feel right.  I recently won a battle just by running my mounted archers away from heavily mounted cavalry.  Not a big deal...but it was in defending my city, just seems odd to be able to kite the enemy.  Plausible strategy sure, but not while receiving the bonuses I get from the fortification upgrades.

There appears to be a melee threat zone you can't pass through which is good (haven't encountered it enough to understand it quite yet).  But this seems to be the only real rule on the field that can alter strategy.  No units can pass though another and no unit can pass by a hostile.

There seems to be an auto-attack?  Hard to catch all the time, but if somehow a path isn't perfect (EG a hostile adjacent) it appears on the last turn they (my guy) will attack no matter what.

Also the marching order frustrates me.  If it was random battle to battle I might be forgiving, but every time my cavalry essentially is behind my archers so I always loose movement and have my soft targets more at risk.

These on top of terrain that doesn't really seem to do anything.  Sometimes it's hard to tell where the legal spots are (brighter grid would help here).   I like the battles being mostly simple, most of the strategy is played in preparation which is fine.  But the battles do seem to hint that there is more going on than there is.  I did finally encounter a map where the center was impassable and the enemy came from around both sides.  I severely outmatched them but the change was nice.