next »

[Tumetsu]

Blue ground may end up being just as common as green ground, but it's an easy change to make, so I'll try it out.

Well, it doesn't need to be blue exactly, but rather something which is more visible. You might get the pop-up effect with grid I already mentioned and perhaps using a simple sprite as as a square with a shading etc. Experimenting probably shows best result.

Also I noticed a few design solutions which I personally would change:

On UI field of character's information, I'd prefer swap the place of health-bar and speed, armor etc. info row. Usually health bars are placed top below character's name and is probably reason why this disturbs me a bit. Also, the placement of "Ice" status breaks the form of the placement of the units in UI. IMO it slightly decreases the readability.
I think you could shorten whole UI area a bit to right, so that healthbar, player text etc. line up with "Ice" text. Alternatively you could space armor, ice etc. line a bit more so that it lines up with rest of the box.

I'd give some more advice but I left my usability course materials to my dorm. They had some very good advice and examples for computer programs. Hmm, looks like I can look lecture pdf's on the net though.

[Soulliard]

I've now had a chance to make many of the UI changes suggested in the last few posts. You can now select attacks from a radial menu when you select a monster, for example.


Now I'm working on adding commanders to the game. This will probably be a good time to explain the spellcasting system.

Scattered across the battlefield are several soul bridges - tiles that represent gateways to the supernal realm. When a monster ends its turn on a soul bridge, it claims the bridge for its team.

At the start of a player's turn, he gains one soul point for each controlled bridge. Commanders can spend soul points to cast spells. Spells are similar to other attacks, but are generally more powerful. For example, they may affect all targets on the battlefield, cause permanent conditions, or change the battlefield's terrain.

The key objective here is to provide multiple routes to victory - going straight for the opposing commander, weakening their minions, or dominating through spells. I want to make it so the winner isn't always obvious until the end of the match. The soul bridge mechanic also encourages creative use of terrain and movement, and discourages completely defensive strategies. I'm looking forward to seeing it all come together.

[Soulliard]

It's about time to introduce a new monster.


Shrike
Sword Spirit

Whenever a truly exceptional weapon is crafted, a shrikes comes into existence. They are initially reclusive, but if they ever come to possess their bonded weapon, they mature into incredible warriors. Their fighting skills are unmatched, since they are literally one with their blade.

Shrikes value nothing more than a challenging fight, regardless of the consequences the fight may have. As long as they are given plenty of opportunities for combat, Shrikes are willing to serve humans of all types, regardless of whether the human is a proper Commander. This makes them popular as elite mercenaries.

Shrikes begin each battle by meditating, which allows them to store up energy for later. They can unleash this energy to move or attack with impossible speed. With enough meditation, a Shrike can take down almost any foe before it even has a chance to react.

Shrikes look remarkably similar to humans, but they are actually composed of soft metal.

[Soulliard]

I've made a lot of improvements since the last demo. Most notably, I've added terrain, commanders, and basic online multiplayer. I've also made some UI improvements and (hopefully) fixed all the bugs that were mentioned earlier in this thread. I haven't tested online multiplayer as much as I would like, so if you try it out, be sure to let me know how it goes.

Download

[Soulliard]

Another new monster:


Pora
Medicine Spirit

Poras are reclusive spirits that hide deep in the forest. Seeing one is very rare, and is considered a symbol of good luck. Stories are told of people who were hopelessly lost in the woods, only to wake up at the edge of the forest with their wounds healed.

Poras hate fighting and lack any sort of combat ability. However, as support units, few can compare. They can heal large amounts of damage, remove any sort of negative condition, and even grant temporary immunity from all attacks.

[BWG]

Looks great. I love this type of game so I will keep an eye on it.
Graphics kind of reminded me of 'Legend' on the amiga.
Keep up the good work

[Soulliard]

Thanks. I never played Legend (it was before my time), but I see the resemblance.

Sorry for the lack of recent updates. I'm at a family gathering, so I don't have much time to work.

[jmcmorris]

The demo is looking great. Monster select can be a little tricky sometimes. Might want to enlarge the hitbox for them. Or maybe I just need to play it more and get used to it.

Are you planning on adding terrain height? That can definitely add a lot of strategy to tactical games. Keep up the good work! I've been a huge fan of the genre since Final Fantasy Tactics - such a good game.

[Soulliard]

Monster select currently requires you to click on the monster's base. I'm going to change it so you can click on any part of the monster.

I don't have any plans to add a complex height system, since I don't want to overcomplicate things. It will probably be mostly aesthetic.

[Soulliard]

It's been a while since I've updated this thread, since I've been busy with family responsibilities. I've mostly been working on the AI anyways, so I haven't had anything terribly exciting to show.

This seems like a good time to introduce the different types of monsters that live in the world.

Elemental monsters are the spirits of raw natural forces, such as lightning, volcanoes and earthquakes. They were birthed during the creation of the world, and their goal is to constantly reshape it (usually destroying what was already there). Their chaotic and destructive attitudes make them frequent enemies of humans and other monsters, but they aren't malicious. They destroy and reshape the land out of passion, not out of hate.

Primal monsters are the spirits of nature. They represent plants, animals, natural landmarks and bodies of water. Their goal is to preserve the world, which frequently puts them at odds with humans and other types of monsters. They are fiercely protective of their land, and they will fight intruders with branch, beak and claw.

Immortal monsters are the spirits of invention. They represent manufactured and constructed things. Their goal is to promote innovation and progress. Unlike other monsters, immortals frequently coexist with humans, since humans are the world's most creative inhabitants. They are integrated into nearly every aspect of society, where they use their powers to improve human life.

Abyssal monsters are the spirits of death. Their ultimate goal is to destroy the world. While some abyssals delight in spreading death and suffering, most find little joy in what they do. All consider death necessary, however, and they will not hesitate to kill when they consider the time to be right.

Supernal monsters are a mystery. They exist in only the most isolated places, and their motivations are unknown. Supernal encounters are extremely dangerous, and they aren't communicative, so few have successfully learned much about them.

[eobet]

Very nice interface! Everything pretty much explains itself.

I think the only thing I'm missing is an indicator that a creature is out of movement points.

[Soulliard]

Sorry for the lack of recent updates. Creating the AI for this game is proving to be a much bigger challenge than I expected. I'm finally at the point where the basic AI is more or less working. Movement is still random, but it selects, targets and orders attacks intelligently.

However, the game hangs for some time while the AI plans its turn. There are some very inefficient parts of the code that I need to refactor, and hopefully that will bring the planning time down to a reasonable level.

If, after the refactoring, the AI still takes an unacceptable amount of time to plan, I'm left with two options:
 - I can reduce the quality of the AI. For example, by having the enemy monsters act in an arbitrary order, rather than in the most efficient order, I can improve performance dramatically (but the AI will not see some good moves).
 - I can recode the project in a lower-level language. This is a nasty scenario, since it means throwing out about 1.5 months of work, but it may be best in the long run.

If you have any advice or heartwarming pictures, I could use them right now. 

[clockwrk_routine]

Instead of arbitrary order, maybe determine by a speed stat.  I'm also working on a tbs sort of, and started on the AI.

I'm thinking, when the playable character finishes moving, a field is calculated around the PC up to the max range of spells, a key:value pair of pc.ID:distanceToPC, would be added to a dictionary contained in a tile, as the field is being calculated. 

While the NPC movement field is being calculated, it should check whether there are values in that dictionary, and if the values are <= to highest range of attack.  If there are it will add those tiles to a list, of potentially targeting characters.  You can pick at random or refine your target by extracting the ID's of the PC also in those dictionaries, and comparing them, who's strongest/weakest/least alive.  Alternatively can have variables referencing the PC that will be dealt the greatest amount of damage, the attack to use, and what tile.

Supporting Allies would work the same way.  You can have separate dictionaries in each tile based on sides, one for friendlies one for enemies.   

Alternatively this might be faster/more intelligent:
You can have another list for characters that are low on health, and characters that have healing abilities.  Characters low on health should reference characters with healing abilities, and move closer to them if they are too far away.  Vice versa, until the character with low health is healed and can be removed from the list.

[Soulliard]

Heh. My avatar is the main character from my last game. He's  a pacifist and generally a pretty average guy, so seeing him go super saiyan is awesome.

Instead of arbitrary order, maybe determine by a speed stat.

I'm thinking something like this would work. It won't be an official attribute, though, since humans will still be able to move monsters in any order they like.

Rather, I'll give monsters with repositioning abilities highest priority in AI planning, since those are most likely to influence the planning of other units. Lowest priority will be given to monsters with mostly debuffing abilities, so they are less likely to waste actions on targets that would die anyways.

[Kinasin]

Looks incredible. 

[Soulliard]

I know the latest updates haven't been too exciting, so here's the details of another monster.


Trolox
Pond Spirit

Trolox dwell at the bottoms of overgrown ponds and swamps, rarely leaving their aquatic homes. Over the years, plants sprout from their bodies, making them difficult to distinguish from the places they inhabit. When they need to interact with the outside world, they command nearby plants to do their bidding.

Their command of plants makes trolox dangerous in battle. They can ensnare foes with thorny vines, making it difficult or impossible for them to move.

[Bandreus]

Pokemon meets Turn based Tactical games? Sign me up on this!

Haven't had the time to check the demo yet, but game looks very nice and the premise is very appealing.

As for the AI: maybe try some simpler/subtler approaches. When you try to make very deep and "smart" AIs, you at time end up with enemy opponents which are pretty much very predictable (cause they always try to do the "best" thing and players might be very good at identifying certain patterns). This might be a welcome result or not.

On the other hand, you could implement some very simple rules for the AI and then let randomization and the current game state determine the kind of actions the AI performs. (Think how algorithms for procedural levels are able to create pretty awesome layouts out of a few basic rules).

It's very hard to give more in-depth help than this without making a TL;DR, so I'll leave that very generic.

Also, I would advice keeping with GM. If it turns out the AI takes up too much CPU time, either optimize it further or make it simpler. You may never know what kind of troubles might pop up when switching to another tech. If you're not happy with GM, you can always put it aside when developing your NEXT game.

Following!

[Paul Jeffries]

I don't know how you're doing the AI now, so this may not be an improvement, but: when I was making my own turn-based tactical game I found Dijkstra maps invaluable.  Essentially from each unit you spread out a set of scores which change with distance to that unit (possibly stopping at some maximum distance).

Computing the map is fairly slow, but to optimise you can keep a separate map for each unit and then just combine them as needed - that way you only need to recalculate the map for a unit when that unit moves.  It also means that you can weight the maps depending on what you're trying to achieve.

So, for the healing example, you could add together the Dijkstra maps of all the units weighted by how desparate they were for health.  The healer could then look at this map for the tiles around them and move 'downhill' to the lowest score (or uphill to the highest one, depending on how you're doing it) until they were within healing range of a unit.  This will give the appearance of the healer prioritizing which unit to help based on how much they need it and also how far away they are (so it won't waste its time trying to help an injured unit the other side of the map when there's another slightly-less-injured unit standing right next to it).

[JobLeonard]

Maybe you can pre-program some small patterns, basically precomputing the AI's tactics?

[Cellusious]

Looks great, when can i trow money at it?