You know, we don't need Deep Blue level of AI skill in Tech battles, would just be fun if the AI used it's abilities and they did something other then pick one guy and chase the hell out of him.  I think some basic tactical battle concepts would have helped WoM, like true ZoC meaning it's difficult to run around units, limited range on spells and bows, flanking, and meaningful terrain.  It has to be hard to really write good AI without those basics in place, I would think.

Sound great. Now tell us how the battles in FE will work...are any of the features Lord Xia mentioned included? Zoc and flanking would be great, range on spells are already in WOM, but different ranges on bows would be a step forward, and maybe spears could have a range of 2 etc. Add terrain features that makes sense (archers on hills!) and we would have a REALLY interesting game combat wise...

And thanks for sharing Frogboy, looking forward to future posts.

Tomas

Cool! Small but interesting update. I'm really looking forward to seeing more of this, since tactical combat is one of the main areas that FE seeks to improve.

Xia, I would argue that those features would make for a more compelling tactical combat experience, but they are in fact additions that would make the combat more complicated, and thus more difficult for the AI to play. I would love to see all of those things, and maybe someday we will get some of them. For now, I would be happy with the first thing you mentioned. A tactical AI that can play the battles as they currently exist. Sounds like Brad is going after things that will make the existing systems fun, stuff like AIs that will use spells intelligently.That is enough of a challenge for now. If they can get all that right, then maybe the next game or expansion in the series can add further depth.

Sounds like an idea for one of them 'Free to Play' games. And a good way for Stardocks studio to increase their exposure.

I admittedly know absolutely nothing about making video games other than the info provided to us on these forums by Stardock. I just think that writing rules on how the computer does things would be easier in a better defined and structured environment.  I am a psychologist, so I am thinking about it as such, a behavior approach.  The less rules the AI has to follow, the less realistic and abnormal their behavior.  When you have a clear and defined structure in which the AI has to function, I would think that would make writing rules for it clearer. Strange paradox about behavior, is that some times when you create limits, you actually create choices. In WoM, there was no way to define a battle formation, everything has near unlimited range, so everything needs to be able to take damage and mitigate damage, so everything needs armor.  Flanking doesn't exist, and neither does true ZoC, and everything counterattacks, so movement and speed are fairly meaningless.  Terrain is all flat and meaningless, so tactics is pretty void.  In a game like that, how can you write any AI other than Shoot and Chase?  There are no other variables, so I would think that would make writing good AI almost impossible.  

Like I said, I think of it as some one trained in behavior, not computers.

Xia, no offense but you don't know what does or does not exist in FE tactical battles. 

And even if you did, what you can and can't do in tactical battles does little to shape the general design structure of writing an AI.

You're right, I only know what existed in WoM, and that is what I was talking about.  But feel free to tell us what FE will have. And like I said, I don't know anything about writing AI.  I'm just a consumer.  Like with a pie, I may not know what goes in a good pie, how to bake it, but I know when one taste like shit.  

We're doing it this way not because modding is a priority but because it makes it a lot easier to write good AI.

For example:

In FE, it's initiative based, not turn based.  So on the left side you get a film strip of what's about to happen (who is about to move).  So a guy on a horse may appear multiple times on that film strip before you get to go.

Anyway, human player chooses to cast Deadly Blizzard which takes about 10 seconds to cast (not "real time" seconds but game seconds). Deadly Blizzard shows up on the film strip above a few other units who get to do their thing first.  One of those units is an AI unit who sees that a human caster is going through the motions to cast deadly blizzard and looks to see if he can counter it.  The Deadly Blizzard spell XML contains a list of suggested counter spells including Warming Aura and that spell is an "instant" (meaning it gets cast instantly) and reduces frost damage by 50%.

Now, without the XML, I'd have to go through the spells in C++ and try to figure out what is a good counter and that decision would be mine alone. Now, back when I wrote the AI and was also the design god, that was possible. But with FE, Derek is the design god so when he (or Toby or whomever) creates a spell they may want to come up with a counter spell too.  It makes it a lot quicker.

Counter-spells sounds awesome... but you know what's even more awesome? AI using it!

Wow...I had NO idea it was going to be like this. This...this is just awesome. Easily one of the most exciting tidbits we have gotten so far.

I rarely ever mess with magic in Fantasy games. I am almost always drawn towards more combative characters, warriors, paladins, etc. In WoM, I often create a Sov with no real magic ability whatsoever! But this...this sounds like so much fun that I am really excited to delve into magic in FE! If you can get the AI to use this system effectively, it will create some AMAZING battles! I can just picture the kick-ass mage duals that will result. While armies clash, two mages battle it out for supremacy with spells and counterspells...really, that is the kind of stuff that we were expecting from War of Magic.

Big thumbs up. Can't wait to hear more, and of course to see it in action.

How far along are you with the AI? I'm assuming not very far, since you just finished 1.4. However, if you are getting into the AI I would imagine that the rest of the game is getting closer to beta, since you previously said you didn't want to get into the AI much while the rest of the game was being worked on, which seems wise since the AI will in large part be based on what is in the rest of the game. I'm not hinting for release dates or beta dates (although I'm insanely eager for both), but I am curious about where your current efforts fit into the bigger development picture.

Thanks!

frogboy, from these comments of yours in this thread, I want the beta to start SOON, as in REAL soon NOW.

harpo

quoted for the truth.. i've been saying this for oh so long about social conventions.. and including games and game design.. so many people point problems in game design down to too many rules disallowing them from building the ultimate 'maths based' combination (an army for example), but in fact rules should be in place to create decision making which by extension the forte of a good strategist..

Some suggestions to make AI design easier which includes ideas from other games:

Chess: Calculate all the possible moves and take the best one. The problem is that there is a lot of randomness in the mechanics which adds a lot of uncertainty making it hard to determine the best move.

R3K6 Method: I am currently playing R3K 6. In that game, battles are managed semi automatically. For each unit, you define a policy and an action. The unit will do the action until complete, then will follow the policy. You can also set additional parameter like unit speed, weapons used, etc. So that could be an idea, design an AI that give orders and design another AI that simply executes them. The only things AI influenced by XML would be the strategist, else the second AI is only path finding and optimisation.

Ogre Battle: For spell and many even unit targetting, Ogre battle allows players to set a proiority (attack strongest, attack weakest, attack leader, etc). When the AI has the choice to target multiple units, the unit selected will be done according to this priority. Ogre battle also have a formation system that dictate how the units will fight. I find this system very interesting and the battles get resolved fast.

Simplify the combat system: Who said that battles needed to be played on a tactical square maps. You could design another system which would be easier to play. For example: you have 3 flanks with a back and front row, you position your units and they attack according to your orders or priorities. It looks similar to ogre battle. The idea is that maneuvering units on a square map makes the AI design much harder. If the battles are easier to manage, so will be the AI design. Ask yourself what kind of decision do you want the player to take in combat and focus on that. A full tactical system sometimes demand the player to do a lot of meaningless uninteresting decision which are not fun. Personally, I would designa system that focus on the fun decisions. Even in a real war, the tactician does not manually move each units, he give them orders.

Many board games try to substitute tactical combat in various ways since they cannot afford to have tactical combat. That is the case for example of the new civilization board game from FFG. There are many combat simulation mechanics out there that can sometime be very creative. Take a look at these mechanics to get some ideas and it would make the combat system redesign. Try it with an paper prototype first. In fact all game mechanics should be tested on paper prototype, even for video games.

Some might say that a full tactical system is essential, but even in Master of magic, sometimes combat just get repetitive and boring to play. There are at most a dozen of combats per game which are actually tense and interesting where strategic decision really matters.  Else, most of the time you auto battle. Tactical battles also takes time, so tactical battles can slowdown the pace of the game especially in multi-player.

Other  suggestions

- Cities should be on the tactical map. there should be gates that players want to defend and capture to enter the city.

- It should be possible to start a battle and include all the surrounding unit stacks of the target on the battlefield, like in age of wonders 2.

- Cloth map mode for tactical battles ... please